It is Adam Smith’s “rent seeking” in a modern hyper-scalable form.
It should be shunned and criminalized on a basis of anti-social economics. Now the other extreme of free software is, arguably, also (economically) bad, as it fails to reward the effort of the creators proportionally to the value gained by the user.
Let us buy our software, and separately, offer us a service agreement that actually has to provide value in its own right. The bundling of these in SaaS is what makes this obscene.
If you are so passionate about it then start a company that builds, deploys and maintains SaaS software on-prem for a comparable cost (say $10/user-mo). If you can provide the same service, guarantees and price points as the typical SaaS while letting companies keep control of their data and preventing lock-in then you'll have the largest market in the world.
And when you had to update version and pay new version license you got push back from bean counters, because current runs fine - and bean counters didn’t like to hear now we have to pay 10x or 100x because no one wanted to spend money on incremental updates because it worked.
You say no need for staff then describe the need for technical staff with specific sysops capabilities.
You can pay for a lot of seats with a $70k salary and that's on the very low end.
If you account for opportunity cost you have to be very large before on-prem becomes worth while or your time must be worth so little that you're probably better doing another business.
But there are tons of small businesses that use IT MSPs. That's how it used to be done for companies that aren't large enough for a full-time IT staff member.
Take for example my friend who works in a ~15 person office in a specialty finance field. Most of their software is SaaS, but they still need someone to manage that software. Plus, the eternal struggle of setting up, managing and supporting employee machines means they need IT services, they just use a local firm. Before SaaS, they would likely have a small on-prem server, but they would be hiring the same MSP firm and probably paying roughly the same amount. SaaS software is normally way more expensive, and that cost increase isn't often offset by a greatly reduced consulting fee for many small businesses.
I have a friend that works at an MSP, they have customers who need help setting up and managing a Square POS system. While that is easier to configure that a locally hosted windows based one from 20 years ago, it's still more than many people opening a restaurant are comfortable with. And is the monthly fee cheaper that a one-time purchase and setup fee plus occasional paid upgrades? I'd genuinely be curious to see an accounting breakdown for a restaurant's software & hardware costs now vs 20 years ago. Hardware is certainly cheaper, but I'm not that convinced that it's really much cheaper overall. Square POS is 69/mo + 50/mo/device for the minimum anything beyond a popup or food truck would need [0], and $828/year minimum isn't exactly free, damn near doubled to $1428 if you have a second register. Especially if you also need to hire the same MSP, just at a potentially reduced number of hours per year.
Start with a few services at ~10$/user/month and it doesn’t take a very large org before the numbers get quite high. And you generally still need some technical support in house.
If you're a large org then adding dedicated staff to manage servers probably makes financial sense but you're also large enough that you're paying support contacts that are probably not far off that Saas price anyhow.
If you're not a larger company you are paying consultants and the licensing fees.
Right but on-prem doesn't mean you escape SaaS pricing. It'll still be $10/user/mo but now you have to run it too. A lot of times you'll pay a premium on the hosted price to get an on-prem version.
Software pricing is a reflection of the fact that it's not a consumable good and the humans that make it have ongoing expenses like rent— not necessarily that it's hosted. Even in the bits in a box days the business was built around recurring revenue, you could choose to not buy the next version but it implicitly relied on most people not doing that.
That depends heavily on how competitive the sector is. Restaurant POS software covers a the largest array of different companies and businesses models I’m aware of. SaaS is generally quite expensive in that market.
In the bits in the box days, quite a lot of the world was skipping version numbers. That was a huge reason for SaaS pricing in the first place. Worse boxed software generally needed some improvement to justify upgrading, SaaS is heavily optimized for rent seeking.
It didn't universally work fine. Every non-technical organization I know of in my circles remembers "the server" with huge disdain, and they're very glad that Google and/or Microsoft has allowed them to get rid of it.
Criticizing harmful practices doesn’t require building a replacement from scratch. Saying “just build your own SaaS company” is a deflection, not a real argument. We don’t demand people create alternative governments to criticize corruption or build new roads to complain about traffic systems. Some practices — like vendor lock-in or exploitative data handling — are worth rejecting simply because they’re harmful.
Also, markets often do come up with better, more ethical alternatives — but only after society pushes back against the harmful defaults. Ecologically, people used to say there was no alternative to burning oil or dumping waste into rivers. That didn’t make those practices okay — and change only came because they were first criticized, regulated, and in some cases outright banned. Progress doesn’t start with “there’s already a perfect alternative”; it starts with “this is no longer acceptable.”
Hmmm.. not everyone is talented in keeping software up to date. A selling homemade earrings can use something like shopify or paypal or stripe easily.
Do you use visa or mastercard? That is SaaS.
The main thing govts need to do is to force companies to allow export/import transparently . But then small SaaS don't have man power to do this (even if they truly believe in cause).
and also penalize or jail both programmrs/CEO that use subscription model with cancellation fee for even cancellations. Lots of people in HN would even be against this - as they claim - I wrote the code only - following orders.
But you'll also compete with very deep VC pockets that don't mind burning lots of money to entrench themselves, so you usually can't compete on cost, until they start the squeeze.
For what its worth, I always feel as if we could provide a binary and then let the binary run on things like aws/gcloud/ preferably cloudflare workers as cf workers feels genuinely good if you focus on typescript/imo sveltekit as well
Please, can anybody enlighten me I suppose? I build a lot of side projects, some of which are genuinely cool and I can genuinely see it being a "saas" but like you, I hate saas from a consumer perspective and I love saas as a guy wanting to earn money so my ethics are definitely questionable at the moment.
Also, let's say I want somebody to buy my software. How does that make sense? What if they just redistribute that software without my permission. Doesn't that just eat into the profits when Saas couldn't.
And also I personally am a foss advocate and as you said its economically bad.
I am kind of thinking of creating software which is just source available and that you need a permission / license which you can get by github sponsoring me but I saw this one project doing it and the people in hackernews were pitch forking the poor guy. Its so crazy and bizarre that once you actually try to provide "some" value. You are automatically the bad guy for not using MIT.
I am also thinking of such a license where its SSPL if you don't have a license or BSD/MIT if you have a license (sort of like redis-ish)
But I am not sure if people would buy them. Maybe I can support both I suppose? I am not sure lol, which is why I wanted to get enlightened
PS: I built a way to create your own cryptocurrencies for genuinely free (no jokes)(not sure how to monetize and if, If anybody can help, that would be great!) and a way to get blogs from youtube posts and a way to youtube community post and videos as an alternative to google photos for unlimited storage (though the privacy aspect for community post is a little bad but I guess I can fix that as well)
Kind of, but not really. I would instead say that the idea of "stickiness" or trying to make your software sticky is rent-seeking. Of course, most SaaS companies vigorously pursue stickiness, but it is not an intrinsic property of SaaS.
Most of that software could run on prem. The problem is that bundling hosting with the actual software in many cases translates into a way to confuse the expense of one for the other. Things that don’t incur operating expenses shouldn’t be charged as a monthly fee, that is rent seeking.
Forcing someone to pay for “services” like hosting, “upgrades”, etc. to gain access to the software is what I find problematic.
Now that the business model has become dominant, it is hard to find any software not on a SaaS-like payment plan. Adobe, Autodesk, Microsoft, etc. all sold software and now all sell subscriptions with little added value to the customer. But the cashflow of a rent model, as forewarned in “Wealth of nations” is so ensnaring it can only be avoided by law. Yet here we are.
But this is still such a bastardization of the term “rent seeking” that is is somewhere between disingenuous and flat out lying.
By this logic, my house cleaner is rent-seeking because I pay every week, but I could do the work myself. That’s not what rent-seeking is. That is a garden variety service.
All of the anti-subscription sentiment just sounds like “I want perpetual support and updates for a one-time price”, which is just silly. It’s actually bad for the customer because once the service provider has your one-time payment, they have zero incentive to keep you as a customer.
It leads to misaligned incentives as sellers seek to expand to new markets to reach new customers while neglecting existing customers who are nothing but expense.
Further, SaaS produces net lower costs because resources can be utilized more efficiently. Great, you can do an on-premise server, but you need to spec it to support the busiest second in the busiest day of your year. Most of the time it will be underutilized.
Sorry, this whole claim is such a massive misunderstanding of Smith and SaaS that it’s making be a bit crazy.
SaaS may not be a particularly good example because like others have mentioned, there are costs to it, and it's up to you to decide whether the price is worth the convenience.
The anti-subscription sentiment is not without merit, however. Software that just runs on your computer now has a fucking subscription for no reason. Adobe, games, etc. That is rent-seeking, because I want to pay for the goddamn thing once and own it. I don't care for support or upgrades; if I do, I'll buy the newer version.
Nobody cares about the costs to offer a product. People care about the value provided.
If I think a service is worth $20 month, that will not change if their cost structure changes and suddenly it costs them $1 or $100 per month to offer. It’s worth $20 to me either way. If their cost becomes $1, I rely on competition to create the consumer surplus.
No, because that would require everyone to own and operate their own servers. I am very happy I do not have to share my bedroom with a server rack so that I can operate my company - not to speak of the cost to deploy a 15gbit/s line to my apartment…
That's just called "over time, some products become worse and more expensive"
Every commercial company is free to charge whatever they want, it's not rent-seeking (unless there is a monopoly, but your argument is applies to small companies as well).
And yes, the fact that you cannot find the software you need for the pricing model you like sucks.. but it is not rent-seeking. And the fact that my local Home Depot does not have cheap, but reliable refrigerators is not rent-seeking either. At worst, it is collusion between manufacturers.
The thing it's being compared to (desktop software) had the same functionality with none of those costs, and it was easier to develop. That's somewhat like a grocery store hiring a dozen egg-sitters to keep the egg supply emotionally supported or some other nonsense and then using that as a justification for the high prices they're able to charge, where an x% profit margin on the unnecessary job translates to significantly more money to the store.
Was desktop software easier to develop than SAAS? It was hard then and is hard now. Even using Electron doesn't remove the need for installers, upgrade paths, logging, diagnostic measures for client installs, troubleshooting when bad hardware is in the mix and only the client can change it.
Doesnt most companies in all industries charge more than their operating costs? Same with people, most work for a salary higher than their operating costs.
Yes, but the point is that rent seeking profits come from things which don’t contribute to economic productivity, which is the accusation being discussed. Having non-zero costs doesn’t imply that you’re not rent seeking.
The problem is that you have really only little competitors on the market.
Eg. Microsoft 365 namely being the one that has outgrown any competitors that can seriously threaten them. They can dictate the prices and they do it willy-nilly and there is no-one that forces them to be cheaper.
Most SaaS contribute to economic productivity though. I see the uptime of local services that our IT maintains, let's just say it's not good. By paying a monthly fee for Slack, we get a working communications system which much better uptime than what we could get from self-hosted.
This is very specific contribution in economic productivity, as the company gets no work done when the communications system is down.
But rent seeking implies some sort of external party enforcing a non-competitive environment that benefits you. Like tariffs or aggressive zoning laws.
Raising the price of your product isn't rent seeking.
You're right, but software does find itself in a non-competitive environment thanks to a sea of intellectual property laws that benefit incumbents. That external party is ever-present, and no doubt implied in discussions here.
Unless you mean copyright / trade-secrets specifically, then IP laws don't really help SaaS that much.
For example take Google Workspace - email, calendar, docs. Google has no particular patents for those, and I've never heard about startups in this area not getting traction because of IP troubles.
Same goes with Jira, or Github, or Slack, or AWS S3 - all of those don't have particularly many patents and in fact there are plenty of alternatives, self-hosted or otherwise. People still pay for them happily.
Nah, rent seeking happens when the customer is faced with little choice (ie: constrained market, can only live there, can't build more, etc...). This is not the case for most of SaaS where you can just signup to another service.
> Now the other extreme of free software is, arguably, also (economically) bad, as it fails to reward the effort of the creators proportionally to the value gained by the user.
What’s the economic explanation then for why so much high quality (or at least, widespread and critical) free software exists?
Probably that free software never actually dies or even really degrades in the traditional sense. Given decades of time, even small incentives to make things better, like the reputational increase one gets from contributing to good FOSS projects, compound really heavily.
Take the Linux kernel as an example. If you were a kernel hacker, even a minor one, from the 1990s, it's quite likely you could parlay that experience into a good job today doing something similar. Those 50-100 hours decades ago have compounded quite nicely for you. But your contribution didn't decay over the next 30 years - worst case scenario, it stayed exactly as good as it was when you stopped, and best case scenario it's been substantially rewritten and incremented upon.
IMO Vendor lock in is when you ask your boss why cant we use tool NEWTHING? Then they tell you its because they have a 5 year contract with Oracle/MS/IBM/Salesforce and that is what we are going to use. We aren't going to have 10 platforms.
Which means in 10 years they will really be locked in because no one is going to un-entrench that thing.
if you've made it to 10 years, that's a really nice problem to have (or maybe a really boring problem to have). But if you're just starting out, I want to prevent you from making all those decisions when you could really be focusing on your startup instead.
So I'm trying to encourage you to consider picking a platform and just sticking with the tools of the platform rather than bundling it yourself together.
It's still a good idea to abstract away these services behind a standardized interface. This way switching from one service to another is just a matter of providing an alternative implementation to said interface.
Granted this approach requires a little foresight...something many companies seem to not have nowadays.
Abstractions are not free either, so if you are creating this "standardized interface", the complexity price you pay is better be worth it.
Often it's less effort to lean in and use all features of the service than to limit yourself to a least common denominator between all competing services.
So the author is saying, don't buy SaaS, that's vendor lock-in. Instead just go all-in on one platform... like Cloudflare, the one (& only) platform that the SDK he writes works on. Which isn't vendor lock-in? No, wait, everything is vendor lock-in:
No matter what choice you make, it's always going to be vendor-locked in.
Switching out something, even if it's open source and self-hosted,
means that you're rewriting a lot of code.
That's not what lock-in means. Just having a vendor-specific component or integration, is not the same thing as being locked-in to a vendor or integration.
Locked-in means that switching it out for something else is either A) impossible, or B) would require an investment greater than just sticking with the existing thing.
When you write software in a loosely-coupled, highly-cohesive way, the intersection between different components is designed to not take much work to replace one component or another. The same is true of systems. If the interfaces of those components are simple, and their use is cohesive, it should not be difficult to replace a part. However, if your components are not cohesive, then it will be a huge pain in the ass to replace anything.
So, no, it's not a good idea to choose a platform because "everything is lock-in, so fuck it, i'll lock myself in even more!" As a developer, I can see the appeal, as it means less work for you. But as a business owner, this is a stupid reason to choose a solution. Choose solutions that will support the business and give it flexibility to change over time.
> But as a business owner, this is a stupid reason to choose a solution. Choose solutions that will support the business and give it flexibility to change over time.
I agree with you. If you're starting out, if your business is not profitable, don't pick SaaS. Don't take the time and pay those 5 taxes. Rather just use the platform, and if you're scalable and profitable and growing, pick some other technology that supports you in the long run.
for what its worth, I use cloudflare workers a lot and I write it in a way where my code can run on anywhere else (and heck, if you really want it, you could run the code locally using wrangler dev I suppose? )
But I guess my software can work with only some changes or I suppose even without some changes on pure node/bun/deno as well
As someone who wants to monetize his side projects, I am not sure what I should do.
Should I make it
1) open source under permissive license (MIT)
2) open source under restrictive license (AGPL/SSPL)
3) source available and only permissive if you pay me a license (like how redis did it in the middle but actually this time , instead of changing license after project is already famous, I do it from the start of the project)
4) not make its source available and distribute a binary for fixed one time.
5) I do any of the above things but with primarily supporting saas? and supporting the ability to move out as you mentioned
Currently, most of my software that I write is just open sourced with MIT and I just private the software that I think has value.
> locking yourself in even more, tying everything to one platform?
I'm saying that if you don't want to use a platform because it's "lock-in," but then use SaaS... then the argument doesn't hold true, especially if you consider the "taxes" of using SaaS.
OP isn't really arguing against SaaS (after all, OP recommends in the end either Cloudflare or Supabase, which are provided as a service...), that's just the clickbait title, rather OP is arguing against signing up for a hundred different vendors and the overhead of commercial relationships with a hundred different vendors.
Which is... not really controversial. Fewer vendors makes your life easier. Fewer dependencies makes your life easier. It would be awesome if you could build your entire product based on the standard library alone! Sadly... that's not really realistic. Nice pipe-dream though.
Thanks! I think you summarized exactly what I was saying, and pointing out that I had a clickbait title. My goal here is if you're starting out with something new, consider reaching for a platform rather than a bunch of services.
The reason why I really love Cloudflare is because of their bindings. A lot of the time you are simply using fetch, so request and response to interact with their services. It feels as if fetch has become like the Unix pipe of the web.
This reads more like a pitch for open-source than anything else.
> Switching out something, even if it's open source and self-hosted, means that you're rewriting a lot of code.
The point of something open-source and self-hosted is that it resolves nearly all of the "taxes" mentioned in the article. What the article refers to as the discovery, sign-up, integration, and local development tax are all easily solved by a good open-source local development story.
The "production tax" (is tax the right word?) can be resolved by contributions or a good plugin/module ecosystem.
With SaaS you mostly give away the chance to benefit from the near zero marginal cost of software. The vendor probably shares the benefit of the marginal cost via lower prices, but at some scale of users and some price per user, the SaaS customer will end up getting a bad deal.
The problem is, you'd be foolish to run your own thing in the early days of a company. It's only when you've succeeded and scaled that it becomes a problem. You survived long enough to need to scale in part by keeping costs low and one way you did that was by using SaaS services instead of building and/or running versions of those tools yourself. That was smart.
As the business grew at least one or two of those SaaS services got so entwined into the daily operations of your company that there is now no way to replace them without a lengthy, risky, and expensive migration project.
The SaaS problem is a negative side effect of your success.
I'm at the stage of my career where vendor lock-in (aka "we are going to use this stable boring corporate-backed tech forever instead of migrating between your hot new JS frameworks every 6 months") is a godsend. Yes I'll happily use AWS. No I will not spend my time to learn and implement RedwoodSDK, whatever that is.
RedwoodSDK is the successor to RedwoodJS. We've rebranded RedwoodJS as "Redwood GraphQL" and built this new thing from scratch. We are the same people with the same ambitions, but more focused. With a narrow niche. I believe that a framework requires a platform to be competitive today.
Because of AI, the difficulty in writing code is greatly reduced. And because of platforms, the difficulty of shipping to production is greatly reduced.
That combination can be really great for your velocity when trying to build a business.
No you didn't rebrand RedwoodJS. You just silently abandoned it. The doc is not even be revised for the proclaimed new name. Paying for vendor to be locked-in into well-funded platform should not be considered as a bad choice after all.
To some degree this happens if you're trying to optimize and find The Best solution for something that isn't core to your product's value proposition. Just take something that meets your requirements and you'll probably be fine. And if you need to switch to something else down the line then, oh well, pull a couple of all-nighters and get it done. Dealing with the ground shifting below you is part of the job.
This is the way to go. Also has the added security bonus of keeping all your services in a private network, as opposed to exposing everything to the internet.
Don't forget that they'll change their API every so often, so you'll spend days and weeks to adapt to "v2" before "v1" is deprecated and eventually removed. You will usually get nothing in terms of desirable features for doing that work, quite often you will get new bugs instead that weren't present with the old API, or even worse, features you depend on are removed because "almost nobody really used them" or "they aren't a good fit in the new interface anymore", and of course you won't have the choice of simply keep using the old version of the thing for an arbitrary amount of time to perform that update on your own pace.
Curious about this. The usual argument is that hosting on Cloudflare is vendor lock-in, I am stating that picking _any_ SaaS induces vendor lock-in.
> No matter what choice you make, it's always going to be vendor-locked in. Switching out something, even if it's open source and self-hosted, means that you're rewriting a lot of code.
The argument is that you might as well not "spend" those 5 taxes, just use the platform, and write the software.
The title strongly implies that vendor lock-in is a bad thing (the phrase "lock-in" has a very much negative connotation), but then the article proposes that you should just give up and go all in on vendor lock-in with a proprietary platform. The alternative to vendor lock-in with SaaS would naturally be running standardized open-source or home grown solutions. That is what people who complain about vendor lock-in generally recommend, not SaaS. The article would be more clear if it addressed that.
Ok, that's fair. I think I might have approached this from a typical JavaScript/ TypeScript developer - where using SaaS is the norm. I'm wondering what you developing in? Not wishing to invalidate your point, just curious?
Consider using SaaS only for non-critical sectors where its discontinuation would not threaten your business's survival. Ensure you have alternatives in place, such as switching to another SaaS provider, self-hosting, or developing an in-house solution. I have witnessed companies fail because their core operations relied too heavily on a SaaS platform."
You're right, I read it a few years ago, and my knowledge is incorrect. I thought the fine was only a percentage of revenue if you're above $10M, or something along those lines.
Allegedly Allscripts (now Veradigm) made their flagship products interoperable, the idea being that customers could choose to buy a giant ecosystem from them, or they could bolt on small solutions onto the giant monoliths that were their competition (like EPIC) and presumably get some revenue they weren't going to get otherwise.
Now - I never was tasked with actually using their software while I worked there, that was just the talking point in the town halls and all hands and such. Being able to export/import was part of that interoperability goal.
Supabase. Big part of why they're heavily adopted. The only issue is that when their client gets big, they usually tend to leave Supabase for more proprietary hosting.
Saas, like streaming media generally, has always been a play to remove the control that comes with ownership from customers and keep it in the hands of big tech. Remember when they discontinued the iPod? It wasn't because people weren't buying.
Apple stopped breaking down iPod sales in 2014, but at that time sales were just 25% of what they were at its peak. By 2022, sales may not have been zero, but it was probably pretty close.
Well SaaS enshittification is among the reasons why open source software is getting more popular. If you are curious why people are sceptical of subscription models look no further than Adobe.
As a former freelancer most of the software I still pay for has a model in the form of: Pay 350 € once, get 2 years of updates and use it as long as you like. If you want new updates after you get a reduced price of 120 € for an extension period.
This is my favourite license model for commercial software since it gives me maximum planability of my expenses and it gives the software company an incentives to add substentially useful features with new updates instead of just collecting rent.
It is Adam Smith’s “rent seeking” in a modern hyper-scalable form. It should be shunned and criminalized on a basis of anti-social economics. Now the other extreme of free software is, arguably, also (economically) bad, as it fails to reward the effort of the creators proportionally to the value gained by the user.
Let us buy our software, and separately, offer us a service agreement that actually has to provide value in its own right. The bundling of these in SaaS is what makes this obscene.
If you are so passionate about it then start a company that builds, deploys and maintains SaaS software on-prem for a comparable cost (say $10/user-mo). If you can provide the same service, guarantees and price points as the typical SaaS while letting companies keep control of their data and preventing lock-in then you'll have the largest market in the world.
It was just called software. People ran it on their own servers. It worked just fine.
and instead of a monthly license you paid an annual support contract plus salaries of staff to maintain your servers and on site support for staff.
And when you had to update version and pay new version license you got push back from bean counters, because current runs fine - and bean counters didn’t like to hear now we have to pay 10x or 100x because no one wanted to spend money on incremental updates because it worked.
SaaS is often aimed at projects with low resources who come for the free tier.
But I can rent a dedicated server at hetzner, use ssh and set it all up. No need for a staff.
The difference is the dedicated server setup requires domain specific knowledge.
If the project needs to scale, a SaaS will become a huge cost in the overall operation and having a dedicated server is much better.
Hiring staff is still cheaper than paying a SaaS, at large scale.
You say no need for staff then describe the need for technical staff with specific sysops capabilities.
You can pay for a lot of seats with a $70k salary and that's on the very low end.
If you account for opportunity cost you have to be very large before on-prem becomes worth while or your time must be worth so little that you're probably better doing another business.
But there are tons of small businesses that use IT MSPs. That's how it used to be done for companies that aren't large enough for a full-time IT staff member.
Take for example my friend who works in a ~15 person office in a specialty finance field. Most of their software is SaaS, but they still need someone to manage that software. Plus, the eternal struggle of setting up, managing and supporting employee machines means they need IT services, they just use a local firm. Before SaaS, they would likely have a small on-prem server, but they would be hiring the same MSP firm and probably paying roughly the same amount. SaaS software is normally way more expensive, and that cost increase isn't often offset by a greatly reduced consulting fee for many small businesses.
I have a friend that works at an MSP, they have customers who need help setting up and managing a Square POS system. While that is easier to configure that a locally hosted windows based one from 20 years ago, it's still more than many people opening a restaurant are comfortable with. And is the monthly fee cheaper that a one-time purchase and setup fee plus occasional paid upgrades? I'd genuinely be curious to see an accounting breakdown for a restaurant's software & hardware costs now vs 20 years ago. Hardware is certainly cheaper, but I'm not that convinced that it's really much cheaper overall. Square POS is 69/mo + 50/mo/device for the minimum anything beyond a popup or food truck would need [0], and $828/year minimum isn't exactly free, damn near doubled to $1428 if you have a second register. Especially if you also need to hire the same MSP, just at a potentially reduced number of hours per year.
[0] https://squareup.com/us/en/point-of-sale/restaurants/pricing
SaaS is rarely cheap by comparison.
Start with a few services at ~10$/user/month and it doesn’t take a very large org before the numbers get quite high. And you generally still need some technical support in house.
If you're a large org then adding dedicated staff to manage servers probably makes financial sense but you're also large enough that you're paying support contacts that are probably not far off that Saas price anyhow.
If you're not a larger company you are paying consultants and the licensing fees.
Right but on-prem doesn't mean you escape SaaS pricing. It'll still be $10/user/mo but now you have to run it too. A lot of times you'll pay a premium on the hosted price to get an on-prem version.
Software pricing is a reflection of the fact that it's not a consumable good and the humans that make it have ongoing expenses like rent— not necessarily that it's hosted. Even in the bits in a box days the business was built around recurring revenue, you could choose to not buy the next version but it implicitly relied on most people not doing that.
That depends heavily on how competitive the sector is. Restaurant POS software covers a the largest array of different companies and businesses models I’m aware of. SaaS is generally quite expensive in that market.
In the bits in the box days, quite a lot of the world was skipping version numbers. That was a huge reason for SaaS pricing in the first place. Worse boxed software generally needed some improvement to justify upgrading, SaaS is heavily optimized for rent seeking.
Until malware (viruses back then) came and data got lost.
Until hard drive died on that server and turns out backups were broken for the last 5 years.
Until someone tried to fix some (real or perceived) problem with the server and it suddenly became full of malware.
It didn't universally work fine. Every non-technical organization I know of in my circles remembers "the server" with huge disdain, and they're very glad that Google and/or Microsoft has allowed them to get rid of it.
We should chat. We're busy building a store for source code.
This would be possible if we could commoditise the operational aspects of cloud computing and SaaS providers. Some companies are trying!
https://onestack.cloud are doing this
Criticizing harmful practices doesn’t require building a replacement from scratch. Saying “just build your own SaaS company” is a deflection, not a real argument. We don’t demand people create alternative governments to criticize corruption or build new roads to complain about traffic systems. Some practices — like vendor lock-in or exploitative data handling — are worth rejecting simply because they’re harmful.
Also, markets often do come up with better, more ethical alternatives — but only after society pushes back against the harmful defaults. Ecologically, people used to say there was no alternative to burning oil or dumping waste into rivers. That didn’t make those practices okay — and change only came because they were first criticized, regulated, and in some cases outright banned. Progress doesn’t start with “there’s already a perfect alternative”; it starts with “this is no longer acceptable.”
Hmmm.. not everyone is talented in keeping software up to date. A selling homemade earrings can use something like shopify or paypal or stripe easily.
Do you use visa or mastercard? That is SaaS.
The main thing govts need to do is to force companies to allow export/import transparently . But then small SaaS don't have man power to do this (even if they truly believe in cause).
and also penalize or jail both programmrs/CEO that use subscription model with cancellation fee for even cancellations. Lots of people in HN would even be against this - as they claim - I wrote the code only - following orders.
But you'll also compete with very deep VC pockets that don't mind burning lots of money to entrench themselves, so you usually can't compete on cost, until they start the squeeze.
For what its worth, I always feel as if we could provide a binary and then let the binary run on things like aws/gcloud/ preferably cloudflare workers as cf workers feels genuinely good if you focus on typescript/imo sveltekit as well
Please, can anybody enlighten me I suppose? I build a lot of side projects, some of which are genuinely cool and I can genuinely see it being a "saas" but like you, I hate saas from a consumer perspective and I love saas as a guy wanting to earn money so my ethics are definitely questionable at the moment.
Also, let's say I want somebody to buy my software. How does that make sense? What if they just redistribute that software without my permission. Doesn't that just eat into the profits when Saas couldn't.
And also I personally am a foss advocate and as you said its economically bad.
I am kind of thinking of creating software which is just source available and that you need a permission / license which you can get by github sponsoring me but I saw this one project doing it and the people in hackernews were pitch forking the poor guy. Its so crazy and bizarre that once you actually try to provide "some" value. You are automatically the bad guy for not using MIT.
I am also thinking of such a license where its SSPL if you don't have a license or BSD/MIT if you have a license (sort of like redis-ish)
But I am not sure if people would buy them. Maybe I can support both I suppose? I am not sure lol, which is why I wanted to get enlightened
PS: I built a way to create your own cryptocurrencies for genuinely free (no jokes)(not sure how to monetize and if, If anybody can help, that would be great!) and a way to get blogs from youtube posts and a way to youtube community post and videos as an alternative to google photos for unlimited storage (though the privacy aspect for community post is a little bad but I guess I can fix that as well)
Kind of, but not really. I would instead say that the idea of "stickiness" or trying to make your software sticky is rent-seeking. Of course, most SaaS companies vigorously pursue stickiness, but it is not an intrinsic property of SaaS.
I don't quite follow this logic. Nearly all SaaS has ongoing costs from the provider: hosting, support, r&d, etc.
How is that rent seeking?
Most of that software could run on prem. The problem is that bundling hosting with the actual software in many cases translates into a way to confuse the expense of one for the other. Things that don’t incur operating expenses shouldn’t be charged as a monthly fee, that is rent seeking. Forcing someone to pay for “services” like hosting, “upgrades”, etc. to gain access to the software is what I find problematic. Now that the business model has become dominant, it is hard to find any software not on a SaaS-like payment plan. Adobe, Autodesk, Microsoft, etc. all sold software and now all sell subscriptions with little added value to the customer. But the cashflow of a rent model, as forewarned in “Wealth of nations” is so ensnaring it can only be avoided by law. Yet here we are.
But this is still such a bastardization of the term “rent seeking” that is is somewhere between disingenuous and flat out lying.
By this logic, my house cleaner is rent-seeking because I pay every week, but I could do the work myself. That’s not what rent-seeking is. That is a garden variety service.
All of the anti-subscription sentiment just sounds like “I want perpetual support and updates for a one-time price”, which is just silly. It’s actually bad for the customer because once the service provider has your one-time payment, they have zero incentive to keep you as a customer.
It leads to misaligned incentives as sellers seek to expand to new markets to reach new customers while neglecting existing customers who are nothing but expense.
Further, SaaS produces net lower costs because resources can be utilized more efficiently. Great, you can do an on-premise server, but you need to spec it to support the busiest second in the busiest day of your year. Most of the time it will be underutilized.
Sorry, this whole claim is such a massive misunderstanding of Smith and SaaS that it’s making be a bit crazy.
SaaS may not be a particularly good example because like others have mentioned, there are costs to it, and it's up to you to decide whether the price is worth the convenience.
The anti-subscription sentiment is not without merit, however. Software that just runs on your computer now has a fucking subscription for no reason. Adobe, games, etc. That is rent-seeking, because I want to pay for the goddamn thing once and own it. I don't care for support or upgrades; if I do, I'll buy the newer version.
Nobody cares about the costs to offer a product. People care about the value provided.
If I think a service is worth $20 month, that will not change if their cost structure changes and suddenly it costs them $1 or $100 per month to offer. It’s worth $20 to me either way. If their cost becomes $1, I rely on competition to create the consumer surplus.
> Most of that software could run on prem.
No, because that would require everyone to own and operate their own servers. I am very happy I do not have to share my bedroom with a server rack so that I can operate my company - not to speak of the cost to deploy a 15gbit/s line to my apartment…
That's just called "over time, some products become worse and more expensive"
Every commercial company is free to charge whatever they want, it's not rent-seeking (unless there is a monopoly, but your argument is applies to small companies as well).
And yes, the fact that you cannot find the software you need for the pricing model you like sucks.. but it is not rent-seeking. And the fact that my local Home Depot does not have cheap, but reliable refrigerators is not rent-seeking either. At worst, it is collusion between manufacturers.
Yeah but who going to fund R&D and stuff like that???
> Most of that software could run on prem.
Many companies don’t have a prem on which to run.
The thing it's being compared to (desktop software) had the same functionality with none of those costs, and it was easier to develop. That's somewhat like a grocery store hiring a dozen egg-sitters to keep the egg supply emotionally supported or some other nonsense and then using that as a justification for the high prices they're able to charge, where an x% profit margin on the unnecessary job translates to significantly more money to the store.
Was desktop software easier to develop than SAAS? It was hard then and is hard now. Even using Electron doesn't remove the need for installers, upgrade paths, logging, diagnostic measures for client installs, troubleshooting when bad hardware is in the mix and only the client can change it.
I kind of wish if we could just slap a desktop software which just runs it on localhost. And that when I click on it, it opens up any of my browsers.
I kind of like this idea personally instead of electron
They presumably charge much more than their ongoing costs. Rent seeking doesn’t mean zero costs.
Doesnt most companies in all industries charge more than their operating costs? Same with people, most work for a salary higher than their operating costs.
Yes, but the point is that rent seeking profits come from things which don’t contribute to economic productivity, which is the accusation being discussed. Having non-zero costs doesn’t imply that you’re not rent seeking.
I know what rent seeking is. That is why it is wrong to claim any salary/profit above operating costs are rent seeking.
So then what is the cutoff and who gets to decide it?
I doubt anything has literally perfectly zero contribution.
The market does. So competitors and demand.
The problem is that you have really only little competitors on the market.
Eg. Microsoft 365 namely being the one that has outgrown any competitors that can seriously threaten them. They can dictate the prices and they do it willy-nilly and there is no-one that forces them to be cheaper.
That's just profit, their incentive to actually build the thing in the first place.
Well yes, rent seeking is specifically profit from things that don’t create new wealth or contribute to economic productivity.
Most SaaS contribute to economic productivity though. I see the uptime of local services that our IT maintains, let's just say it's not good. By paying a monthly fee for Slack, we get a working communications system which much better uptime than what we could get from self-hosted.
This is very specific contribution in economic productivity, as the company gets no work done when the communications system is down.
I sell my saas to willing buyers at the fair market price.
That’s not mutually exclusive with rent seeking, of course.
But rent seeking implies some sort of external party enforcing a non-competitive environment that benefits you. Like tariffs or aggressive zoning laws.
Raising the price of your product isn't rent seeking.
You're right, but software does find itself in a non-competitive environment thanks to a sea of intellectual property laws that benefit incumbents. That external party is ever-present, and no doubt implied in discussions here.
Unless you mean copyright / trade-secrets specifically, then IP laws don't really help SaaS that much.
For example take Google Workspace - email, calendar, docs. Google has no particular patents for those, and I've never heard about startups in this area not getting traction because of IP troubles.
Same goes with Jira, or Github, or Slack, or AWS S3 - all of those don't have particularly many patents and in fact there are plenty of alternatives, self-hosted or otherwise. People still pay for them happily.
Nah, rent seeking happens when the customer is faced with little choice (ie: constrained market, can only live there, can't build more, etc...). This is not the case for most of SaaS where you can just signup to another service.
Exercising available pricing power isnt the same thing as rent seeking.
> Now the other extreme of free software is, arguably, also (economically) bad, as it fails to reward the effort of the creators proportionally to the value gained by the user.
What’s the economic explanation then for why so much high quality (or at least, widespread and critical) free software exists?
Probably that free software never actually dies or even really degrades in the traditional sense. Given decades of time, even small incentives to make things better, like the reputational increase one gets from contributing to good FOSS projects, compound really heavily.
Take the Linux kernel as an example. If you were a kernel hacker, even a minor one, from the 1990s, it's quite likely you could parlay that experience into a good job today doing something similar. Those 50-100 hours decades ago have compounded quite nicely for you. But your contribution didn't decay over the next 30 years - worst case scenario, it stayed exactly as good as it was when you stopped, and best case scenario it's been substantially rewritten and incremented upon.
That's how I explain it to myself at least.
That's only true for libraries because developers benefit from bugs getting fixed. FLOSS applications are far from high quality.
This is exactly how I'm thinking of monetizing the framework.
so basically s/subscription/service agreement/ ?
IMO Vendor lock in is when you ask your boss why cant we use tool NEWTHING? Then they tell you its because they have a 5 year contract with Oracle/MS/IBM/Salesforce and that is what we are going to use. We aren't going to have 10 platforms.
Which means in 10 years they will really be locked in because no one is going to un-entrench that thing.
if you've made it to 10 years, that's a really nice problem to have (or maybe a really boring problem to have). But if you're just starting out, I want to prevent you from making all those decisions when you could really be focusing on your startup instead.
So I'm trying to encourage you to consider picking a platform and just sticking with the tools of the platform rather than bundling it yourself together.
You can stick with it without legally getting trapped by some shady BD getting their sales quota filled
It's still a good idea to abstract away these services behind a standardized interface. This way switching from one service to another is just a matter of providing an alternative implementation to said interface.
Granted this approach requires a little foresight...something many companies seem to not have nowadays.
The problem is people don’t just store data on systems like Salesforce, they use them to build very complex applications.
Getting your data into and out of Salesforce is easy, it has excellent APIs. Rewriting your applications is the bigger hurdle.
Abstractions are not free either, so if you are creating this "standardized interface", the complexity price you pay is better be worth it.
Often it's less effort to lean in and use all features of the service than to limit yourself to a least common denominator between all competing services.
So the author is saying, don't buy SaaS, that's vendor lock-in. Instead just go all-in on one platform... like Cloudflare, the one (& only) platform that the SDK he writes works on. Which isn't vendor lock-in? No, wait, everything is vendor lock-in:
That's not what lock-in means. Just having a vendor-specific component or integration, is not the same thing as being locked-in to a vendor or integration.Locked-in means that switching it out for something else is either A) impossible, or B) would require an investment greater than just sticking with the existing thing.
When you write software in a loosely-coupled, highly-cohesive way, the intersection between different components is designed to not take much work to replace one component or another. The same is true of systems. If the interfaces of those components are simple, and their use is cohesive, it should not be difficult to replace a part. However, if your components are not cohesive, then it will be a huge pain in the ass to replace anything.
So, no, it's not a good idea to choose a platform because "everything is lock-in, so fuck it, i'll lock myself in even more!" As a developer, I can see the appeal, as it means less work for you. But as a business owner, this is a stupid reason to choose a solution. Choose solutions that will support the business and give it flexibility to change over time.
> But as a business owner, this is a stupid reason to choose a solution. Choose solutions that will support the business and give it flexibility to change over time.
I agree with you. If you're starting out, if your business is not profitable, don't pick SaaS. Don't take the time and pay those 5 taxes. Rather just use the platform, and if you're scalable and profitable and growing, pick some other technology that supports you in the long run.
for what its worth, I use cloudflare workers a lot and I write it in a way where my code can run on anywhere else (and heck, if you really want it, you could run the code locally using wrangler dev I suppose? )
But I guess my software can work with only some changes or I suppose even without some changes on pure node/bun/deno as well
Perhaps the difference between a religion and a cult is that you can leave a religion.
Likewise .. if you can get your data in a standard format and walk away, you are not locked-in.
Customers tend to feel less aggrieved when they have access to their data - too many SaaS platforms dont allow this.
Really interesting.
As someone who wants to monetize his side projects, I am not sure what I should do.
Should I make it 1) open source under permissive license (MIT)
2) open source under restrictive license (AGPL/SSPL)
3) source available and only permissive if you pay me a license (like how redis did it in the middle but actually this time , instead of changing license after project is already famous, I do it from the start of the project)
4) not make its source available and distribute a binary for fixed one time.
5) I do any of the above things but with primarily supporting saas? and supporting the ability to move out as you mentioned
Currently, most of my software that I write is just open sourced with MIT and I just private the software that I think has value.
So to fight vendor lock in.. double down by locking yourself in even more, tying everything to one platform?
> locking yourself in even more, tying everything to one platform?
I'm saying that if you don't want to use a platform because it's "lock-in," but then use SaaS... then the argument doesn't hold true, especially if you consider the "taxes" of using SaaS.
OP isn't really arguing against SaaS (after all, OP recommends in the end either Cloudflare or Supabase, which are provided as a service...), that's just the clickbait title, rather OP is arguing against signing up for a hundred different vendors and the overhead of commercial relationships with a hundred different vendors.
Which is... not really controversial. Fewer vendors makes your life easier. Fewer dependencies makes your life easier. It would be awesome if you could build your entire product based on the standard library alone! Sadly... that's not really realistic. Nice pipe-dream though.
Thanks! I think you summarized exactly what I was saying, and pointing out that I had a clickbait title. My goal here is if you're starting out with something new, consider reaching for a platform rather than a bunch of services.
The reason why I really love Cloudflare is because of their bindings. A lot of the time you are simply using fetch, so request and response to interact with their services. It feels as if fetch has become like the Unix pipe of the web.
This reads more like a pitch for open-source than anything else.
> Switching out something, even if it's open source and self-hosted, means that you're rewriting a lot of code.
The point of something open-source and self-hosted is that it resolves nearly all of the "taxes" mentioned in the article. What the article refers to as the discovery, sign-up, integration, and local development tax are all easily solved by a good open-source local development story.
The "production tax" (is tax the right word?) can be resolved by contributions or a good plugin/module ecosystem.
Open source is free if your time is worth nothing.
some people just don't understand business
people is gonna find out why companies pays top dollar for close source alternative vs open source product
With SaaS you mostly give away the chance to benefit from the near zero marginal cost of software. The vendor probably shares the benefit of the marginal cost via lower prices, but at some scale of users and some price per user, the SaaS customer will end up getting a bad deal.
The problem is, you'd be foolish to run your own thing in the early days of a company. It's only when you've succeeded and scaled that it becomes a problem. You survived long enough to need to scale in part by keeping costs low and one way you did that was by using SaaS services instead of building and/or running versions of those tools yourself. That was smart.
As the business grew at least one or two of those SaaS services got so entwined into the daily operations of your company that there is now no way to replace them without a lengthy, risky, and expensive migration project.
The SaaS problem is a negative side effect of your success.
edited - a typo and a word change
I'm at the stage of my career where vendor lock-in (aka "we are going to use this stable boring corporate-backed tech forever instead of migrating between your hot new JS frameworks every 6 months") is a godsend. Yes I'll happily use AWS. No I will not spend my time to learn and implement RedwoodSDK, whatever that is.
There's a RedwoodJS frontend framework, never heard of RedwoodSDK
Nobody ever got fired for hiring IBM
RedwoodSDK is the successor to RedwoodJS. We've rebranded RedwoodJS as "Redwood GraphQL" and built this new thing from scratch. We are the same people with the same ambitions, but more focused. With a narrow niche. I believe that a framework requires a platform to be competitive today.
Because of AI, the difficulty in writing code is greatly reduced. And because of platforms, the difficulty of shipping to production is greatly reduced.
That combination can be really great for your velocity when trying to build a business.
No you didn't rebrand RedwoodJS. You just silently abandoned it. The doc is not even be revised for the proclaimed new name. Paying for vendor to be locked-in into well-funded platform should not be considered as a bad choice after all.
I can’t tell if serious or not, but I threw up in my mouth a little regardless.
Dead serious.
I’m now similarly conflicted between wanting to say congratulations or condolences.
Thanks, that does make sense, I just hadn't heard about the new direction/rebranding.
Agreed on both of your assesments Best of luck!
Thank you! We're getting a lot of love! Just need to solve distribution. ;P
That's why there's so much of it. Annuity like income and pricing power is an attractive business model to create around.
The discovery tax is actually something like O(n log n), because you have to also search and also compare. It's part of The Mess We're In[0].
0. https://www.youtube.com/watch?v=lKXe3HUG2l4
To some degree this happens if you're trying to optimize and find The Best solution for something that isn't core to your product's value proposition. Just take something that meets your requirements and you'll probably be fine. And if you need to switch to something else down the line then, oh well, pull a couple of all-nighters and get it done. Dealing with the ground shifting below you is part of the job.
SaaS is the solution to get devs to actually pay for tooling, like other professionals.
Who doesn't like it, should promote that upstream gets more than bug reports and push requests, as means to pay their bills.
Use only open-source SaaS with the option to self-host.
This is the way to go. Also has the added security bonus of keeping all your services in a private network, as opposed to exposing everything to the internet.
The real issues is that, 3 years down the line, they'll dramatically the price, and then raise the price again, and again.
Because that's the incentive, particularly with products that are naturally fading and ceasing to make new sales.
Don't forget that they'll change their API every so often, so you'll spend days and weeks to adapt to "v2" before "v1" is deprecated and eventually removed. You will usually get nothing in terms of desirable features for doing that work, quite often you will get new bugs instead that weren't present with the old API, or even worse, features you depend on are removed because "almost nobody really used them" or "they aren't a good fit in the new interface anymore", and of course you won't have the choice of simply keep using the old version of the thing for an arbitrary amount of time to perform that update on your own pace.
Good point, I think I should add that to the article.
The important thing about vendor lock in is allowing another company to have pricing leverage over you.
The title is incongruous with the conclusion.
Curious about this. The usual argument is that hosting on Cloudflare is vendor lock-in, I am stating that picking _any_ SaaS induces vendor lock-in.
> No matter what choice you make, it's always going to be vendor-locked in. Switching out something, even if it's open source and self-hosted, means that you're rewriting a lot of code.
The argument is that you might as well not "spend" those 5 taxes, just use the platform, and write the software.
The title strongly implies that vendor lock-in is a bad thing (the phrase "lock-in" has a very much negative connotation), but then the article proposes that you should just give up and go all in on vendor lock-in with a proprietary platform. The alternative to vendor lock-in with SaaS would naturally be running standardized open-source or home grown solutions. That is what people who complain about vendor lock-in generally recommend, not SaaS. The article would be more clear if it addressed that.
Ok, that's fair. I think I might have approached this from a typical JavaScript/ TypeScript developer - where using SaaS is the norm. I'm wondering what you developing in? Not wishing to invalidate your point, just curious?
We're a Rails shop generally using basic Postgres + Redis with various Rails gems filling in where needed. Our needs aren't terribly complex though.
Consider using SaaS only for non-critical sectors where its discontinuation would not threaten your business's survival. Ensure you have alternatives in place, such as switching to another SaaS provider, self-hosting, or developing an in-house solution. I have witnessed companies fail because their core operations relied too heavily on a SaaS platform."
Not if it’s open data and people can easily move their data between vendors
That's the dream right, how often is that true? Seems like this should be a leading indicator for picking a SaaS!
Finding just the right way to tell management it's time to re-architect the monolith and move into infra-as-code to become agnostic is the dream.
This is a legal requirement of GDPR, no? https://gdpr-info.eu/art-20-gdpr/
Edit: I am wrong.
AFIAK GDPR only applies if you're profitable, otherwise a fine on revenue ... pre-revenue isn't applicable.
I don't see anything about revenue here https://commission.europa.eu/law/law-topic/data-protection/r...
You're right, I read it a few years ago, and my knowledge is incorrect. I thought the fine was only a percentage of revenue if you're above $10M, or something along those lines.
The GDPR always applies regardless of whether a company is profitable or not. But it covers only personal data/sensible data, not "all data".
Not once have I ever seen this happen
Allegedly Allscripts (now Veradigm) made their flagship products interoperable, the idea being that customers could choose to buy a giant ecosystem from them, or they could bolt on small solutions onto the giant monoliths that were their competition (like EPIC) and presumably get some revenue they weren't going to get otherwise.
Now - I never was tasked with actually using their software while I worked there, that was just the talking point in the town halls and all hands and such. Being able to export/import was part of that interoperability goal.
It's so rare, and apparently so unimportant, that I've never seen it as part of the marketing strategy.
Supabase. Big part of why they're heavily adopted. The only issue is that when their client gets big, they usually tend to leave Supabase for more proprietary hosting.
Saas, like streaming media generally, has always been a play to remove the control that comes with ownership from customers and keep it in the hands of big tech. Remember when they discontinued the iPod? It wasn't because people weren't buying.
Apple stopped breaking down iPod sales in 2014, but at that time sales were just 25% of what they were at its peak. By 2022, sales may not have been zero, but it was probably pretty close.
Well SaaS enshittification is among the reasons why open source software is getting more popular. If you are curious why people are sceptical of subscription models look no further than Adobe.
As a former freelancer most of the software I still pay for has a model in the form of: Pay 350 € once, get 2 years of updates and use it as long as you like. If you want new updates after you get a reduced price of 120 € for an extension period.
This is my favourite license model for commercial software since it gives me maximum planability of my expenses and it gives the software company an incentives to add substentially useful features with new updates instead of just collecting rent.
Ooh, are we just saying nonsense phrases now? How about “React is just race conditions with better branding”